*Clean revised Manuscript Click here to view linked References
1 Effects of flushing flows on the transport of mercury-polluted particulate matter
2 from the Flix Reservoir to the Ebro Estuary
3 Albert Palanquesa,*, Jorge Guilléna, Pere Puiga, Joan O. Grimaltb
4 a Institute of Marine Sciences (CSIC), Passeig Maritim de la Barceloneta, 37-49,
6 b Institute of Environmental Assessment and Water Research (CSIC), Jordi Girona, 18,
7 Barcelona 08034, Spain
8 *Corresponding author. Tel.: +34 932309500; fax: +34 932309555. E-mail address:
9 [email protected] (A. Palanques).
10
11
12
13
14
15
16
17
18
19
20
21
1
22 Abstract
23
24 Polluted sediments retained in water reservoirs are potential sources of
25 deleterious materials downstream, especially during floods or flushing flows (FFs).
26 Their interaction with these events is important for determining potential risks and
27 evaluating management actions. In the Ebro River, the Flix Reservoir accumulated a
28 deposit of more than 3 x105 t of industrial waste with high Hg concentrations. Because
29 suspended particulate matter (SPM) is the main driver of Hg pollution downriver, this
30 study analyses the transport of particulate matter and Hg pollution from the Flix
31 Reservoir to the Ebro Estuary during FFs. Time series of currents, turbidity and
32 downward particulate matter fluxes were obtained by current meters, turbidimeters and
33 sediment traps assembled in benthic tripods. They were deployed in the reservoir and at
34 two locations in the estuary during two recording periods that each captured a flushing
35 flow (FF) event. In addition, SPM samples were collected during the study period at
36 several locations along the river course, from upstream of the Flix Reservoir down to
37 the river mouth, to measure the suspended particulate matter and associated Hg
38 mobilized downstream.
39 A continuous background level of Hg pollution was observed during the
40 deployment periods, but the Hg and particulate matter fluxes increased by between one
41 and two orders of magnitude during FFs. Though the two events reached similar water
42 discharges, the first FF was after the wet season and generated lower particulate matter
43 concentrations and fluxes, but higher Hg contents than the second, which occurred after
44 the dry season. The higher available particulate matter in the second event diluted the
45 polluted Hg particle load more than the first event. Thus, similar FFs may result in
2
46 different Hg concentration and sediment transport episodes, largely depending on the
47 previous hydrological regime and the river sediment availability. These findings should
48 be considered for FF management.
49
50 Keywords: flushing flow, mercury transport, particulate matter, polluted reservoir, salt
51 wedge estuary, Ebro River.
52
53 1. Introduction
54
55 The uncontrolled dumping of waste and the construction of dams in many rivers
56 around the world have favoured the accumulation of anthropogenic deposits in fluvial
57 systems on a global scale (Wildi et al., 2004; Syvitski et al., 2005; García Bravo et al.,
58 2009; Man et al., 2013; Skalak et al., 2013). With about 800,000 dams around the
59 world, more than 50% of the river sediment flux is trapped in regulated watersheds
60 (Vörösmarty et al., 2003; Nilsson, 2009), and water reservoirs are on the whole an
61 effective global pollution trap (Wildi et al., 2004). These historical deposits are a
62 pollution source that could affect the quality of water used for irrigation and human
63 consumption. The more erodible part of these deposits can leak from the reservoir along
64 with the lighter fine suspended sediment and be transported downstream. In addition,
65 environmental management may induce changes in physicochemical conditions, leading
66 to the release of pollutants (Frémion et al., 2016)
67 Many pollutants are linked to fine particulate matter (<63 µm) (Feng et al.,
68 1998; Israelsson et al 2014) that can bypass the dams and flow downstream to be
3
69 trapped in estuaries due to the convergence of flows associated with their circulation
70 (Yellen et al., 2017; Ralston and Geyer, 2017; Burchard et al., 2018). There, fine
71 particles and pollutants are recirculated and can be periodically deposited and
72 resuspended (Garnier et al., 1991; Grabemann et al., 1997; Turner and Millward, 2002;
73 Geyer and Ralston, 2018).
74 Under the Mediterranean climate, long periods of low river flow during droughts
75 alternate with flash floods, and most of the river sediment transport takes place during
76 flood events (Batalla et al., 1995; García et al, 2000; Vericat and Batalla, 2006). During
77 the 20th century, many dams were built for water management and hydroelectric energy
78 production, leading to artificial flood management. Controlled flood flows from dams,
79 known as FFs, are used for environmental and/or engineering purposes (Kondolf and
80 Wilcock, 1996; Batalla et al., 2006; Batalla and Vericat, 2009).
81 In the lower Ebro River (Fig. 1), FFs have been implemented since 2003, mainly
82 to remove the excess of macrophytes, and these events have been monitored, analysed
83 and modelled (Batalla and Vericat, 2009; Tena et al., 2012). FFs have also been
84 suggested as a sediment management tool for increasing the sediment delivered to the
85 coast and delta plain (Rovira and Ibáñez, 2007). During FFs, the SPM concentration
86 doubles that of natural floods, although discharges are typically lower due to their
87 shorter peaks and duration (Tena et al., 2012).
88 The last reservoir of the Ebro River (the Flix reservoir) is highly polluted by the
89 mercury (Hg) of a waste deposit retained on it since 1948 (Palanques et al., 2014). The
90 effect of FFs on the transport of Hg to the Ebro Estuary has not been addressed up to
91 now. This transport is closely tied to the transport of particulate matter because of the
92 great affinity of Hg for particles (Gunnarson et al., 2009; Masson et al., 2018).
4
93 In this paper, we study the effects of two FFs on the particulate matter and Hg
94 transport from the Hg-polluted Flix Reservoir to the Ebro Estuary. Like the Flix
95 Reservoir, many dams in the world were originally built for industrial purposes and now
96 retain polluted sediment (Lenhart, 2003). The interaction of these sediments with river
97 floods and FFs is important for determining their potential risks and evaluating
98 management actions. Moreover, though many studies have characterized contaminants
99 in steady conditions or in sediment records, fewer have considered the transport
100 dynamics of suspended sediment–borne pollution (Quesada et al., 2014; Geyer and
101 Ralston, 2018).
102
103 2. Study area and Methods
104
105 2.1 Study area
106
107 The Ebro River, the second largest river flowing into the NW Mediterranean,
108 drains an area of around 85,000 km2 that was progressively impounded during the 20th
109 century and now has close to 190 dams (Vericat and Batalla, 2006). The largest
110 complex of reservoirs in the Ebro River is formed by the Mequinenza (water capacity:
111 1534 hm3), Riba-roja (water capacity: 207 hm3) and Flix (water capacity: 11 hm3) dams.
112 This complex is located in the lower reaches of the river and impounds 97% of the basin
113 (Fig. 1A). The suspended sediment discharge in the pre-dam period was estimated to be
114 between 18 and 21 x 106 t y-1, whereas in the post-dam period it has only been about 0.1
115 x 106 t y-1 (Palanques et al., 1990; Guillén and Palanques, 1992; Vericat and Batalla,
116 2006).
5
117 The last downstream dam, that of the Flix Reservoir, is located 110 km upstream
118 from the Ebro River mouth. The Flix dam was built in 1948 only a few hundred metres
119 downstream from an industrial complex located on the south river bank (Fig. 1B). In
120 1949, this industrial complex was transformed into a chlor-alkali plant using the
121 mercury cathode electrolytic technique, which requires large amounts of mercury and is
122 well-known for its environmental impact (UNEP, United Nations Environment
123 Programme, 2002; OSPAR Commission, 2005; Ullrich et al., 2007). As a consequence,
124 the Flix Reservoir and the lower Ebro River have been affected by the residue dumped
125 from the plant for decades. This residue formed an industrial waste deposit of more than
126 3 x 105 t that obstructed about 50% of the reservoir water section (Fig. 1C), and Hg is
127 the trace element that shows the highest anomalies, reaching concentrations of up to 640
128 mg kg−1 (4 orders of magnitude higher than natural levels in river mud) (Palanques et al.
129 2014). The Hg pollution from the Flix industrial complex affects organisms living in the
130 Flix Reservoir and downstream, such as worms (Ramos et al. 1999), zebra mussels
131 (Carrasco et al. 2008), crayfish (Suarez-Serrano et al. 2010), carps (Navarro et al. 2009)
132 and catfish (Carrasco et al. 2011).
133 Between the Flix dam and the Ebro Delta plain, the river is gravel-bedded
134 (Vericat and Batalla 2006), and polluted suspended particles flow towards the estuary.
135 There is no significant particulate Hg pollution source downstream from the Flix
136 complex to the estuary. The Ebro River has a microtidal salt wedge estuary, and the salt
137 wedge can intrude more than 30 km when the water discharge decreases below 150-200
138 m3 s-1. Between this discharge and 350-400 m3 s-1, the head of the salt wedge is
139 generally located at Gracia Island, 15 km from the river mouth, which acts as a
140 topographic threshold for the intrusion. Above about 400 m3 s-1, the salt wedge is
141 pushed out of the river (Guillén and Palanques, 1992).
6
142 Dam removal or failure can result in decades of accumulated material being
143 released downstream rapidly and catastrophically (Stanley and Dole 2003). This risk is
144 now being eliminated because the Flix waste deposit has been isolated from the main
145 river course, and decontamination works to remove the waste from the reservoir are
146 underway (Palanques et al., 2014). The data shown in this study were obtained before it
147 was decided to remove the waste deposit from the Flix Reservoir, and the results will be
148 important reference parameters for evaluating the evolution of Hg pollution in the lower
149 Ebro River when the total decontamination of the reservoir has been completed.
150
151 2.2. Field work
152
153 River particulate matter is retained in the reservoir and the estuary, where it
154 settles. Downward particulate matter was collected by vertical sediment traps to
155 determine its pollutant concentrations and estimate the trapped downward particle
156 fluxes. In addition, as SPM is the main driver of Hg pollution downriver, time series of
157 currents and turbidity data were recorded to assess the transport of particulate matter.
158 Three instrumented benthic tripods were deployed during two one-month
159 periods in which FFs took place. Deployment 1 was from 12 April to 12 May 2006 and
160 Deployment 2 from 11 November to 11 December 2006. One tripod was deployed in
161 the Flix Reservoir at 3.5 m depth and the other two in the Ebro Estuary near Gracia
162 Island and near the river mouth at 5.50 m depth (Fig. 1A). Each tripod was equipped
163 with an Aanderaa Doppler current meter (RCM-9) with temperature, salinity, 0-25
164 formazin turbidity unit (FTU) and 0-500 FTU range turbidity sensors, and also with a
165 vertical cylindrical sediment trap (100 cm high and 15.2 cm wide) with 10 receiving
7
166 cups. The current, temperature, salinity and turbidity sensors were located 60 cm above
167 the bottom, and the top of the trap was located 150 cm above the bottom. The current
168 meter sampling intervals were set at 5 min and the sediment trap collecting interval was
169 set at 3 days for each cup.
170 The sediment trap cups were filled with a borax-buffered 5% formaldehyde
171 solution in 0.20 μm filtered seawater before their deployment. This poisoning solution
172 limits degradation of trapped particles and prevents the mechanical disruption of
173 swimming organisms (“swimmers”) that occasionally enter the traps (e.g., Knap et al.,
174 1996). After recovery, the cups were stored in the dark at 2–4 °C until they could be
175 processed in the laboratory.
176 During the four field campaigns for the two deployments and the two recoveries
177 of the benthic tripods carried out on 11 April, 13 May, 10 November and 12 December,
178 a set of vertical hydrographic profiles were recorded using a Sea-Bird SBE 9 CTD
179 coupled with a Seapoint turbidimeter. Water samples were taken with a Niskin bottle
180 attached to the CTD in surface and near-bottom waters at the tripod sites. Water was
181 also sampled upstream from the Flix Reservoir, near the town of Riba-roja, as a control
182 station to determine the Hg levels of the SPM reaching the Flix Reservoir (Fig. 1A), and
183 along the north and the south banks of the Flix Reservoir (see location in Fig 1C). The
184 November FF was also monitored by taking water samples every 1-2 hours at Riba-roja,
185 at the Flix meander (just after the reservoir), at Mora d’Ebre and at Gracia Island in the
186 estuary (see location in Fig. 1A). Water samples were stored in the dark at 2–4 °C until
187 they could be processed in the laboratory
188 The Confederación Hidrográfica del Ebro provided hourly daily water discharge
189 at the gauging stations of Ascó and Tortosa (see location in Fig. 1A).
8
190
191 2.3 Laboratory work
192
193 Settling trapped material was split with a peristaltic dispenser to divide the total
194 sample into several homogeneous aliquots. “Swimmers” were removed from the
195 samples by wet-sieving the sample through a 1 mm nylon mesh. Samples were left to
196 settle again (for one to two days) and examined under a magnifying glass to remove the
197 “swimmer” organisms smaller than 1 mm. By this method, the bulk of swimmers were
198 removed and the “biological” contamination, if any, was negligible (Heussner et al.,
199 2006). Subsequently, the samples were washed with Milli-Q water and centrifuged three
200 times to extract all the remaining seawater. Finally, the samples were freeze-dried by
201 lyophilisation and weighed. They were then homogenized in an agate mortar for the
202 analysis of Hg. Downward total mass flux (TMF) was computed using the total mass
2 203 weight divided by the trap collecting area (182 cm ) and the sampling interval (3 days).
204 Water samples were filtered onto Whatmann glass microfibre filters (GFFs) to
205 obtain the SPM concentration and suspended sediment samples for analysis of their Hg
206 content. Before use, the GFFs were rinsed with Milli-Q water, placed for 24 hours in an
207 oven at 550º Celsius, allowed to cool for another 24 h, and then weighed. After filtering
208 of the water samples, the filters with the suspended sediment samples were dried to
209 constant weight at 40º Celsius and then placed in a desiccation bowl for 24 h. The SPM
210 concentration was estimated by dividing the weight of the suspended sediment by the
211 volume of filtered water in each sample.
9
212 The turbidity data from the tripods and the CTDs were recorded in FTU and
213 converted into estimates of SPM concentration through regression curves from paired
214 measurements of turbidity and SPM concentration obtained with water samples taken
215 with the Niskin bottle coupled with the CTDs at the tripod sites. In the estuary the
216 calibration curve was
217 SPM concentration = 0.94 FTU-1.99 (R² = 0.66)
218 and in the reservoir it was
219 SPM concentration = 2.05 FTU-12.32 (R2=0.57).
220 As the regression curves were not highly significant and turbidity-SPM
221 concentration rating curves can vary in time, reported values from the time series should
222 be considered estimates of SPM concentration in order to visualize the temporal
223 evolution of this parameter throughout the deployment period.
224 The Hg content of the downward and suspended particulate matter samples was
225 analysed using a LECO AMA254 Mercury Analyzer complying with the US EPA
226 Method 7473 (US EPA, 2007). The instrument’s detection limit of Hg is 0.01 ng.
227 PACS-2 certified reference material from the National Research Council Canada and
228 blank samples were used for analytical quality control. Thirty-five replicate
229 determinations of the certified reference material with a certified value of 3.04 ± 2.0 mg
230 kg-1 (mean and SD) gave an average value of 2.98 ± 0.1 mg kg-1. Downward Hg fluxes
231 were obtained by multiplying downward TMF by the Hg concentrations of the trapped
232 particle samples.
233
10
234 3. Results
235
236 3.1. River water discharge
237
238 River water discharge during the two FFs reached similar maximum values:
239 1498 m3 s-1 in May and 1539 m3 s-1 in November at the Ascó gauging station and 968
240 m3 s-1 in May and 895 m3 s-1 in November at the Tortosa gauging station. The lower
241 peak at the Tortosa station was a consequence of the change in the shape of the
242 hydrograph due to the floodwave attenuation downstream. The duration of these events
243 was also similar, in both cases about 15 h. In the rest of the deployment periods, the
244 river discharge ranged between 120 and 200 m3 s-1 (Fig. 2).
245 In this paper, because the Ascó gauging station is closer to the Flix Reservoir
246 and the Mora d’Ebre sampling sites and Tortosa is the closest gauging station to the
247 estuary (Fig. 1A), we will refer to the Ascó water discharge for the Flix and Mora
248 d’Ebre data and to the Tortosa water discharge for the estuary data.
249 The context in which the two FFs occurred was different. The first FF took place
250 one month after a period of high water discharge (>400 m3 s-1) and a natural flood (up to
251 1530 m3 s-1), whereas the second FF took place after six months of low water discharge
252 (mainly <200 m3 s-1) (Fig. 2).
253
254 3.2. Hg concentration in SPM along the lower Ebro River
255
11
256 Hg concentrations in the SPM upstream of the Flix Reservoir at the Riba-roja
257 station ranged between 0.07 and 0.43 mg kg-1, with a mean value of 0.19 mg kg-1 (Table
258 1; Fig. 3). These are the Hg levels of the SPM reaching the Flix Reservoir, which are
259 relatively low. At the south bank of the Flix Reservoir, over the waste deposit, the Hg
260 levels of the SPM were between two and four orders of magnitude higher than those at
261 Riba-roja (Table 1; Fig. 3). On the north bank, however, where the riverbed is not
262 covered by the industrial waste, the Hg levels were only one order of magnitude higher
263 than upstream of the Flix Reservoir (Fig. 3; Table 1), suggesting that the particles
264 leaving the waste deposit probably reach the deeper channelized part of the reservoir
265 (Figs. 1C) and are mixed with the upstream inputs, diluting the pollution concentration.
266 Downstream of the Flix dam, Hg concentrations were similar to those on the
267 north bank of the reservoir (Fig.3; Table 1). In the estuary, Hg values decreased slightly,
268 but still maintained Hg concentrations about one order of magnitude higher than
269 upstream of the reservoir (Fig. 3; Table 1), possibly indicating that the polluted particles
270 escaping from the waste deposit and mixed with the upstream inputs bypass the dam
271 and are transferred directly downstream to the estuary.
272 In the estuary, an upper fresh water layer up to 2 m thick and a sharp horizontal
273 gradient to the salt water below were observed during the four field campaigns to install
274 and recover the tripods. At Gracia Island, the SPM and Hg concentrations tended to be
275 slightly higher in the surface fresh water than in the bottom salt water, whereas at the
276 river mouth, surface and near-bottom values were more similar (Supplementary
277 Material Fig. 1).
278 During the November 21 FF, Hg concentrations of SPM upstream of the
279 reservoir maintained relatively constant values (around 0.16 mg kg-1). However,
12
280 downstream of the reservoir the concentrations increased to 2.40 mg kg-1 at the Flix
281 meander, to 2.25 mg kg-1 at Mora d’Ebre and to 0.80 mg kg-1 at Gracia Island (Fig. 4).
282 These increases lasted only two hours and took place at the beginning of the FF, when
283 more erodible polluted sediment was available for the increasing currents, and before
284 the SPM concentration and water discharge peaks. The Hg concentration decreased
285 when the SPM concentration increased at the peak of the flow, when most sediment
286 removal from bed and bank erosion occurred along the river, thus diluting the pollution
287 load.
288 Discharge of SPM-borne pollution can be estimated where gauging stations
289 measure the volumetric flow rate of water that is transported through a given river
290 cross-sectional area, assuming homogeneous pollutant and SPM concentrations across
291 it. In the case of the November FF event, Hg discharge can be roughly estimated
292 extrapolating the water discharge from the Ascó gauging station and from the SPM and
293 Hg concentrations monitored at Riba-roja and Mora d’Ebre during this event
294 (Supplementary Material Fig. 2). The maximum Hg discharge was 18 mg s-1 at Riba-
295 roja (upstream of the reservoir), and 672 mg s-1 at Mora d’Ebre (downstream of the
296 reservoir).
297 At Gracia Island, the salinity record indicates that the fresh river water occupied
298 the whole estuary section during the November FF event, so the Hg discharge was
299 roughly estimated in the same way, but the water discharge was extrapolated from the
300 Tortosa gauging station, giving a maximum Hg discharge of 156 mg s-1 (Supplementary
301 Material Fig. 2).
302 From the SPM samples taken when the river discharge was between 120 and 200
303 m3 s-1, we can also estimate Hg discharges ranging from 0.2 to 0.4 mg s-1 at Riba-roja
13
304 and from 0.8 to 2.4 mg s-1 at Mora d’Ebre. All these data confirm the downriver
305 transport of Hg from the reservoir, especially during FFs.
306
307 3.3. Time series of hydrodynamics, SPM concentration and particle matter fluxes
308
309 3.3.1. The Flix Reservoir
310 In the Flix Reservoir, daily oscillations of current speed (from 1 to 15 cm s-1)
311 and SPM concentrations (from 1 to 30 mg L-1) were recorded during most of the
312 deployment periods, and were probably induced by the dam’s water management (Fig.
313 5). However, current speed and SPM concentrations peaked to 23 cm s-1 and 411 mg L-1
314 in the May FF and to 37 cm s-1 and 980 mg L-1 in the November FF, respectively.
315 Downward TMF was between 6 and 27 g m-2 d-1 during most of deployment 1 and
316 between 50 and 100 g m-2 d-1 during most of deployment 2, increasing to 105 g m-2 d-1
317 in the sample affected by the May FF and to 390 g m-2 d-1 in the sample affected by the
318 November FF (Fig. 5). Therefore, SPM concentration and downward TMF were two to
319 four times higher during the November FF than during the May FF. Recorded peaks of
320 SPM concentrations and downward particle fluxes corresponded mostly to polluted
321 sediment resuspended at the reservoir during the FFs.
322
323 3.3.2. The Ebro Estuary
324 During most of the two deployment periods, time series at the estuary stations
325 were recorded in the salt wedge intruding the estuary, where near-bottom currents
326 ranged between 1 and 15 cm s-1, alternating from up-estuary (~270º) to down-estuary
14
327 (~90º) (Figs. 6 and 7). Under these conditions, the SPM concentration ranged between 1
328 and 12 mg L-1 and downward TMF between 5 and 40 g m-2 d-1. However, during the
329 May and November FFs, current speed at Gracia Island increased to over 50 cm s-1,
330 flowing seawards and pushing the salt water and the sediment retained in the estuary
331 downriver for 1-2 days (Fig. 6). At this site, SPM concentration and downward TMF
332 increased by between one and two orders of magnitude during the FFs, and these values
333 were around twice as high in the November FF as in the May FF (Fig. 6, and Table 2).
334 At the river mouth, the FFs induced lower peaks of near-bottom currents (32 and
335 38 cm s-1) and short periods (only up to six hours) of near-bottom downriver flow of
336 brackish waters (salinity of 24.06 in spring and 9.52 in autumn) (Fig. 7). The near-
337 bottom SPM concentration and downward TMF did not increase during the May FF,
338 and the SPM concentration increased very sharply to 184 mg L-1 during the November
339 FF, but just for two hours.
340
341 3.4. Time series of Hg concentrations and fluxes of the trap-collected particulate matter
342
343 Hg concentrations of the downward particulate matter at the Flix Reservoir site
344 were between two and three orders of magnitude higher than natural values and were
345 higher in the first deployment than in the second (Fig. 8). The FFs resuspended polluted
346 reservoir sediment and increased the downward Hg fluxes about four times, reaching
347 2020 and 4860 µg m-2 d-1 in the sampling periods coinciding with the May and
348 November FFs, respectively (Fig. 8).
349 In the Ebro Estuary, the downward Hg concentration during the two deployment
350 periods were almost one order of magnitude higher than natural values and were higher
15
351 in the first deployment than in the second. (Fig. 8). At Gracia Island, the FF events
352 increased the downward Hg fluxes by more than one order of magnitude in the
353 corresponding trap samples, reaching a similar value in both FFs (Fig. 2; Table 2). By
354 contrast, there were no increases in the downward Hg fluxes linked to the FFs at the
355 river mouth (Fig. 8).
356 The sediment trap samples showed no clear increase in the Hg concentration
357 induced by the FFs (Fig. 8). The two-hour increase detected during the November FF
358 monitoring (Fig. 4) was probably masked in the sediment trap samples because they
359 include three-day periods of downward particles.
360
361 4. Discussion
362
363 4.1 Comparison of the May and November FFs
364
365 Sediment dynamics in regulated rivers is a function not only of the hydrology
366 during the events, but also of several parameters arising from the preceding conditions,
367 such as sediment availability, exhaustion and flashiness, and loss of energy due to flow
368 routing (Tena et al., 2012). In the Ebro River, the May and November FFs sharply
369 increased current speed, SPM concentration and downward particle and Hg fluxes (Figs.
370 5, 6, 7, 8). These events were similar in water discharge and duration, but the SPM
371 concentration and downward TMF were lower and the Hg concentration higher during
372 the May FF than during the November FF (Figs 5, 6, 7, 8; Supplementary Material Fig.
373 3; Table 2). Batalla and Vericat (2009) measured a maximum SPM concentration of 363
16
374 mg L-1 during the November FF and of only 88 mg L-1 during the May FF, with mean
375 values of 244 and 56.7 mg L-1, respectively, thus corroborating the lower sediment
376 transport during the May FF.
377 The different sediment transport during the two FFs could be explained by the
378 river conditions before each event (Fig. 2). The May FF occurred after the wet season.
379 The previous natural floods could have transported the more erodible sediment
380 downstream, winnowing the river and the reservoir, pushing back the salt wedge and
381 leaving a lower fine sediment load available to be transported along the river and
382 estuary during the first deployment and FF. This reduction of the sediment availability
383 probably decreased the dilution of the Flix polluted particles with upstream particles,
384 giving higher Hg concentrations in May than in November (Fig. 8, Tables 2 and 3).
385 By contrast, the November FF was in the dry season after six months of low
386 water discharge (< 200 m3 s-1), which allowed a higher sediment load to be stranded in
387 the river, in the reservoir and in the estuary, where the salt wedge intrusion is
388 maintained at discharges below about 400 m3 s-1 (Guillén and Palanques 1992). All this
389 would have left more sediment available to be transported during the November FF than
390 during the May FF. The higher availability of low-polluted particles from upstream of
391 the Flix Reservoir in November may have diluted and decreased the Hg concentrations
392 of the trap-collected particulate matter more than in May (Fig. 8; Tables 2 and 3). One
393 fact that supports this hypothesis is that downward TMF in the May FF at Gracia Island
394 was approximately half that in the November FF (Fig. 6; Table 2), but Hg
395 concentrations in the May event were almost double those in the November event, so
396 the resultant downward Hg flux at Gracia Island was similar during the two events (Fig.
397 8, Table 2).
17
398
399 4.2 Effects of FF in the particulate matter and Hg transport of the estuary
400
401 In estuaries, currents and SPM concentrations can change strongly with depth
402 and also across and along them according to the interactions between marine and
403 freshwater. Therefore, the estimate of the total SPM and Hg discharges requires data
404 about the vertical and lateral structure of the velocity and SPM distributions, which are
405 difficult to obtain. However, an approach can be made by estimating the integrated
406 sediment transport calculated from the time series of current velocity and SPM
407 concentrations at a single location, which provides an indication of the temporal
408 variability of sediment transport at a single point (Geyer and Ralston, 2018). Assuming
409 that particles move with the velocity of the water within which they are suspended
410 (Wright, 1995), the instantaneous near-bottom sediment flux (q) in g m-2 s-1 at the
411 height of the instrument is obtained as the product of the velocity module c and the
412 SSC, in mg/L:
413 q(t) = c (t)SSC (t)
414 The integration of the instantaneous horizontal fluxes for the duration of the
415 deployments yielded the cumulative horizontal suspended sediment transport in kg m-2.
416 The cumulative horizontal SPM transport at the estuary site of Gracia Island
417 indicates that there was a dominant up-river inflow of SPM in the salty layer during the
418 salt wedge intrusion periods (104 and 293 kg m-2 during D1 and D2 respectively;
419 Supplementary Material Fig. 3). This helped retain fine sediment and associated Hg
420 inside the estuary during low water discharge periods, leaving them available to be
421 transported during FFs and natural floods. By contrast, the cumulative horizontal SPM
18
422 transport during the FFs was down-estuary and increased by between one and two
423 orders of magnitude, and twice more in the November event than in the May event
424 (3764 and 9535 kg m-2 respectively; Supplementary Material Fig. 3 and Table 2).
425 Another rough estimation of the sediment transport in the estuary can be made
426 from the SPM concentration recorded by the Gracia Island turbidimeter and the water
427 discharge from the Tortosa gauging station during the FFs, when the river water
428 occupied the whole estuary section. The SPM discharges estimated in this way show a
429 maximum value in the November FF (352 kg s-1) almost twice that in the May FF (190
430 kg s-1).
431 An approach to estimate the cumulative suspended Hg transport can be made
432 from the cumulative SPM transport with the assumption that the Hg concentration of the
433 trap samples collected during the FFs are the mean Hg concentration of the SPM during
434 these events. These calculations give a Hg transport increase of two orders of magnitude
435 during both FFs (from about 0.06 g m-2 to a maximum of 3.61 g m-2 in the May FF and
436 3.99 g m-2 in the November FF).
437 Following the same assumption for the mean Hg concentrations during the FFs,
438 another approach can be made to estimate the Hg discharge during these events,
439 considering the SPM concentration recorded by Gracia Island turbidimeter and the
440 water discharge from the Tortosa gauging station. These calculations give similar
441 maximum Hg discharges of 173 g s-1 in the May FF and 166 g s-1 in the November FFs.
442
443 4.3. Implications of the FFs in the estuary
444
19
445 Although these SPM and Hg transport estimates have uncertainties, they suggest
446 that similar maximum suspended Hg fluxes and discharges were reached during both
447 FF´s, as the downward Hg fluxes also did (Fig. 8). However, the higher sediment load
448 transported during the November event diluted more the Hg concentration of the
449 particulate matter transported to the estuary than the May FF. This has implications for
450 the assessment of toxicity risks. Long et al. (1995) defined two values for Hg to
451 establish three ranges according to which the effects to the environment are predicted to
452 be minimal (value < 1.5 mg kg-1), occasional (1.5 mg kg-1 value < 0.71 mg kg-1) or
453 frequent (value > 0.71 mg kg-1). In November, Hg concentrations were within the range
454 of occasional effects but in May they were within the range of frequent effects. This
455 kind of effects should be considered for FF management.
456 The absence of major suspended and downward particle flux increases at the
457 river mouth station during the FFs (Fig. 7; Supplementary Material Fig. 3; Table XX)
458 suggests that the sediment outflow at this site was mainly through the upper freshwater
459 layer. This phenomenon may be favoured by the downstream thinning of the freshwater
460 layer and the higher flow speed in this layer (Guillén and Palanques 1992), which
461 prevents the mixing of particles between fresh and salt layers.
462 FFs eject downstream the salt water at the head of the estuary (near Gracia
463 Island), but in the two cases studied they did not last long enough to push out the salt
464 water from the river mouth. This could also mean that part of the sediment transported
465 by the FFs did not reach the river mouth, remaining in the estuary until new, larger
466 floods transferred it to the marine environment. This is important for managing river
467 sediment transport, because it could indicate that stronger or longer flows must be
468 generated if the purpose of the FF is not only to remove the excess of macrophytes but
469 also to transfer river sediment into the coastal environment.
20
470 Hg discharges from chlor-alkali plants have also spread into other river and
471 estuarine systems such as the Alveiro Lagoon, the Mersey Estuary and the Penobbscot
472 Estuary. Hg concentrations in suspended sediment of the Alveiro Lagoon ranged
473 between 0.46 and 6.7 mg kg-1 at the mouth of the channel discharging from the plant
474 and between 0.33 and 1.7 mg kg-1 about 1-2 km down-estuary (Pereira et al., 1998). In
475 the Mersey Estuary, the Hg concentration of the surface bottom sediment was between
476 0.6 and 1.0 mg kg-1 (Harland et al., 2000) and in the Penobscot Estuary it was between
477 0.12 and 2.75 mg kg-1 (Merritt and Amirbahman, 2007; Yeager et al., 2018). The Ebro
478 estuary has a lower tidal range than the above estuaries and, unlike them, it has a dam
479 downstream from the pollution source.
480 Increases in sediment transport and pollutant load during FFs and natural floods
481 have also been recorded in the Soča River in the Gulf of Trieste. This river was affected
482 by the Hg pollution from a former mine in Idrija, and a flood of similar characteristics
483 to the studied ones (1600 m3 s-1) increased the Hg concentration of SPM to 46.6 mg kg-1
484 (Rajar et al., 2000). This is one order of magnitude higher than downriver from the Flix
485 reservoir, probably because, independently of the pollution sources, the Soča River had
486 no dam downstream of the mine. The Flix dam, which was built just downriver of the
487 chlor-alkali plant, retained the industrial waste in the reservoir and limited the release of
488 polluted particles downstream. This dam reduced the transport of the polluted load
489 towards the estuary, discharging it episodically or leaking it in continuous lower fluxes.
490 The Hg concentrations detected in the Ebro Estuary were similar to those
491 detected in the Rhone Delta during the 1980s, ranging mainly from 0.6 to 0.9 mg kg-1
492 (Figueres et al., 1985, Cossa and Martin, 1991), but they were significantly higher than
493 those detected at the Beaucaire gauging station (located 60 km upstream from the
494 Rhone River mouth) between 2011 and 2016, which showed a mean Hg concentration
21
495 of 0.095 mg kg-1 (Poulier, 2019). This means that, over the past 30 years, major
496 improvements in wastewater treatment through implementation of the Urban
497 Wastewater Treatment Directive (UWWTD) have helped reduce the pollutants in the
498 Rhone River (Olivier et al., 2008), whereas in the Ebro River, the industrial waste
499 deposit of the Flix Reservoir is a historic Hg source that still needs to be removed.
500 A significant decrease in the Hg pollution in the Ebro Estuary is expected after
501 the decontamination of the Flix Reservoir. However, in some estuaries the bidirectional
502 transport of sediments and the estuarine regime leads to a long dilution time scale, as
503 occurred in the Penobscot Estuary, where mercury released into the river in the 1970s
504 now has a nearly uniform concentration in mobile SPM and a residence time of
505 approximately 25 years. (Geyer & Ralston 2018).
506
507 5. Conclusions
508
509 This study indicates that polluted sediments accumulated in reservoirs can
510 become a permanent source of pollutants downstream and can either be discharged
511 episodically or leak continuously in lower fluxes towards estuarine systems, where they
512 are recirculated.
513 The FFs in the lower Ebro River sharply increased the SPM concentration, the
514 particulate matter fluxes and the Hg fluxes in the reservoir and in the estuary at Gracia
515 Island. The Hg concentration also increased but only during the first 2 hours of the
516 November FF. Most of the downriver particulate matter and Hg transport in the estuary
517 at Gracia Island occurred during the FFs.
22
518 Similar FF events can produce different effects in SPM and particulate Hg
519 transport, depending on the previous dynamics of the river system. After a long low
520 water discharge period, the November FF generated higher particulate matter fluxes
521 than after a high river discharge period in the May FF, mainly due to higher sediment
522 availability in the November event. However, the Hg concentration was lower during
523 the November FF because the higher available particulate matter diluted the polluted Hg
524 particle load more than during the May event and the resultant Hg fluxes in both events
525 were similar.
526 The reduction of the Hg concentration when there is more particulate matter
527 availability in the river to dilute the pollution load should be considered for FF
528 management. In addition, this study gives reference parameters for controlling the
529 evolution of Hg pollution in the lower Ebro River after the decontamination of the
530 reservoir, in order to determine how the system responds to the elimination of this
531 pollution source.
532
533 Acknowledgements
534
535 This study was financed by the Spanish Ministry of Environment and the
536 Catalan Water Agency (ACA) through the MOBITROF Project (Estudio de la
537 movilidad de los metales pesados, compuestos organoclorados y radionúclidos del
538 Embalse de Flix y de su acumulación en las cadenas tróficas). A.P., J.G. and P.P. belong
539 to the CRG on Littoral and Oceanic Processes, supported by grant 2017 SGR 863.
540
23
541 References
542
543 Batalla, R.J., Sala, M., Werritty, A., 1995. Sediment budget focused in solid material
544 transport in a subhumid Mediterranean drainage-basin. Z. Geomorphol. 39, 249–64.
545
546 Batalla, R.J., Vericat, D., Palau, A., 2006. Sediment transport during a flushing flow in
547 the lower Ebro River, in: Rowan, J.S., Duck, R.W., Werritty, A. (Eds), Sediment
548 Dynamics and the Hydromorphology of Fluvial Systems. IAHS Publication 306,
549 Wallingford, pp. 37–44.
550
551 Batalla, R.J., Vericat, D., 2009. Hydrological and sediment transport dynamics of
552 flushing flows: implications for management in large Mediterranean rivers. River Res.
553 Appl. 25, 297–314.
554
555 Burchard, H., Schuttelaars, H.M., Ralston, D.K., 2018 Sediment Trapping in Estuaries.
556 Annu. Rev. Mar. Sci. 10, 371395.
557
558 Carrasco, L., Díez, S., Soto, D.X., Catalan, J., Bayona, J.M., 2008. Assessment of
559 mercury and methylmercury pollution with zebra mussel (Dreissena polymorpha) in the
560 Ebro River (NE Spain) impacted by industrial hazardous dumps. Sci. Total Environ.
561 407, 178–84.
562
24
563 Carrasco, L., Benejam, L., Benito, J., Bayona, J.M., Díez, S., 2011. Methylmercury
564 levels and bioaccumulation in the aquatic food web of a highly mercury-contaminated
565 reservoir. Environ. Int. 37, 1213–1218.
566
567 Cossa, D., Martin, J.M., 1991. Mercury in the Rhône delta and adjacent marine areas.
568 Mar. Chem. 36 (1–4), 291–302.
569
570 Feng, H., Cochran, J.K., Lwiza, H., Brownawell, B.J., Hirschberg, D.J., 1998.
571 Distribution of heavy metal and PCB contaminants in the sediments of an urban estuary:
572 the Hudson River. Mar. Environ. Res. 45 (1), 69–88.
573
574 Figueres, G., Martin, J.M., Meybeck, M. and Seyler, P., 1985. A comparative study of
575 rnercury contamination in the Tagus Estuary (Portugal) and major French estuaries
576 (Gironde, Loire, Rhone). Estuar. Coast. Shelf. S. 20, 183-203.
577
578 Frémion, F., Bordas, F., Mourier, B., Lenain, J.F, Kestens, T., Courtin-Nomade, A.,
579 2016. Influence of dams on sediment continuity: A study case of a natural metallic
580 contamination. Sci. Total Environ, 547, 282-294.
581
582 García, C., Laronne, J.B., Sala, M., 2000. Continuous monitoring of bedload flux in a
583 mountain gravel-bed river. Geomorphology 34, 23–31.
584
25
585 García Bravo, A., Loizeau, J.L., Ancey, L., Ungureanu, V.G., Dominik, J., 2009.
586 Historical record of mercury contamination in sediments from the Babeni Reservoir in
587 the Olt River, Romania. Environ. Sci. Pollut. Res. Int. 16 (Suppl. 1), S66–75.
588
589 Garnier, J.M., Martin, J.M., Mouchel, J.M., Thomas, A. J., 1991. Surface reactivity of
590 the Rhône suspended matter and relation with trace element sorption. Mar. Chem. 36,
591 267–289.
592
593 Geyer, W.R., Ralston, D.K., 2018. A mobile pool of contaminated sediment in the
594 Penobscot Estuary, Maine, USA. Sci. Total Environ. 612, 694–707
595
596 Grabemann, I., Uncles, R.J., Krause, G., Stephens, J.A., 1997. Behaviour of turbidity
597 maxima in the Tamar (UK) and Weser (FRG) estuaries. Estuar. Coast. Shelf Sci. 45 (2),
598 235–246
599
600 Guillén, J., Palanques, A., 1992. Sediment dynamics and hydrodynamics in the lower
601 course of a river highly regulated by dams: the Ebro River. Sedimentology 39, 567-579.
602
603 Gunnarson, S.J., Granberg, M.E., Nilsson, H.C., Rosenberg, R., Hellman, B., 2009.
604 Influence of sediment organic matter quality on growth and polychlorobiphenyl
605 bioavailability in Echinodermata (Amphiura filiformis). Environ. Toxicol. Chem. 18
606 (7), 1534–1543.
26
607
608 Harland, B.J., Taylor, D., Wither, A., 2000. The distribution of mercury and other trace
609 metals in the sediments of the Mersey Estuary over 25 years 1974-1998. Sci. Total
610 Environ. 253, 45-62.
611
612 Heussner, S., Durrieu de Madron, X., Calafat, A., Canals, M., Carbonne, J., Delsaut,
613 N.,Saragoni, G., 2006. Spatial and temporal variability of downward particle fluxes on a
614 continental slope: lessons from an 8-yr experiment in the Gulf of Lions (NW
615 Mediterranean). Mar. Geol. 234, 63–92.
616
617 Israelsson, P.H., Quadrini, J.D., Connolly, J.P., 2014. Fate and transport of hydrophobic
618 organic chemicals in the lower Passaic River: insights from 2, 3, 7, 8-
619 tetrachlorodibenzo-p-dioxin. Estuar. Coasts 37, 1145–1168.
620
621 Kondolf, G.M., Wilcock, P.R., 1996. The flushing flow problem: defining and
622 evaluating objectives. Water Resour. Res. 32(8), 2589–2599.
623
624 Knap, A., Michaels, A., Close, A., Ducklow, H. Dickson, A. (eds.), 1996. Protocols for
625 the Joint Global Ocean Flux Study (JGOFS) Core Measurements. JGOFS Report Nr.
626 19, vi+170 pp. Reprint of the IOC Manuals and Guides No. 29, UNESCO 1994.
627
628 Lenhart, C.F., 2003. An assessment of NOAA community-based fish passage and dam
629 removal projects. Coast. Manage. 231, 77–96.
27
630
631 Long, E.R., Macdonald, D.D., Smith, S.L., Calder, F.D., 1995. Incidence of adverse
632 biological effects within ranges of chemical concentrations in marine and estuarine
633 sediments. Environ. Manag. 19, 81–97.
634
635 Merritt, K.A., Amirbahman, A., 2007. Mercury dynamics in sulfide-rich sediments:
636 geochemical influence on contaminant mobilization within the Penobscot River estuary,
637 Maine, USA. Geochim. Cosmochim. Acta 71, 717–722.
638
639 Masson, M., Angot, H., Le Bescond, C., Launay, M., Dabrin, A., Miège, C., Le Coz, J.,
640 Coquery, M., 2018. Sampling of suspended particulate matter using particle traps in the
641 Rhône River: relevance and representativeness for the monitoring of contaminants.
642 Sci.Total Environ. 637–638, 538–549 VSI: Rhône Sediment
643
644 Man, K.C., Peck, J.A., Peck, M.C., 2013. Assessing dam pool sediment for
645 understanding past, present and future watershed dynamics: An example from the
646 Cuyahoga River, Ohio. Antropocene 2, 76–88.
647
648 Navarro, A., Quiros, L., Casado, M., Faria, M., Carrasco, L., Benejam, L., et al., 2009.
649 Physiological responses to mercury in feral carp populations inhabiting the low Ebro
650 River (NE Spain), a historically contaminated site. Aquat. Toxicol. 93, 150–7.
651
28
652 Nilsson, C., 2009. Reservoirs. Encyclopedia of Inland Waters. Elsevier, pp. 625–33.
653
654 Olivier, J.M., Dole-Olivier, M.J., Amoros, C., Carrel, G., Malard, F., Lamouroux, N.,
655 Bravard, J. P., 2008. The Rhone River Basin. In: Tockner, K., Robinson, C.T.,
656 Uehlinger, U. (Eds). Rivers of Europe. Academic Press. pp. 247-295.
657
658 OSPAR Commission 2005. Hazardous Substances Series: Mercury losses from the
659 chlor-alkali industry in 2003.
660 http://www.ospar.org/documents/dbase/publications/p00600/p00600_mercury_losses_r
661 eport_2003.pdf
662
663 Palanques, A., Plana, F., Maldonado, A., 1990. Recent influence of man on the Ebro
664 Margin Sedimentation System, Northwestern Mediterranean. Mar. Geol. 95, 247–63.
665
666 Palanques, A., Grimalt, J., Belzunces, M., Estrada, F., Puig, P., Guillén, J., 2014.
667 Massive accumulation of highly polluted sedimentary deposits by river damming. Sci.
668 Total Environ. 497–498, 69–381
669
670 Pereira, M.E., Duarte, A.C., Millward, G.E., Vale, C., Abreu, S.N., 1998. Tidal export
671 of particulate mercury from the most contaminated area of Aveiro’s Lagoon, Portugal.
672 Sci. Total Environ. 213, 157-163
673
29
674 Poulier, G., Launay, M., Le Bescond, Ch., Thollet, F., Coquery, M., Le Coz, J. 2019.
675 Combining flux monitoring and data reconstruction to establish annual budgets of
676 suspended particulate matter, mercury and PCB in the Rhône River from Lake Geneva
677 to the Mediterranean Sea. Sci. Total Environ. 658, 457-473.
678
679 Quesada, S., Tena, A., Guillén, D., Ginebreda, A., Vericat, D., Martínez, E., Navarro-
680 Ortega, A., Batalla, R.J., Barceló. D., 2014. Dynamics of suspended sediment borne
681 persistent organic pollutants in a large regulated Mediterranean river (Ebro, NE Spain).
682 Sci. Total Environ. 473–474, 381-390.
683
684 Rajar, R., Žagar, D., Širca, A., Horvat, M., 2000. Three-dimensional modelling of
685 mercury cycling in the Gulf of Trieste. Sci. Total Environ. 260, 109-123.
686
687 Ralston, D.K., Geyer, W.R., 2017. Sediment transport timescales and trapping
688 efficiency in a tidal river J. Geophys. Res-Earth 122, 2042-2063.
689
690 Ramos, L., Fernandez, M.A., Gonzalez, M.J., Hernandez. L.M., 1999. Heavy metal
691 pollution in water, sediments, and earthworms from the Ebro River, Spain. Bull.
692 Environ. Contam. Toxicol. 63, 305–11.
693
694 Rovira, A., Ibáñez, C., 2007. Sediment Management Options for the Lower Ebro River
695 and its Delta. J Soils Sediments 7 (5), 285–295.
30
696
697 Skalak, K.J., Benthem, A.J., Schenk, E.R., Hupp, C.R., Galloway, J.M., Nustad, R.A.,
698 Wiche, G.J., 2013. Large dams and alluvial rivers in the Anthropocene: the impacts of
699 the Garrison and Oahe Dams on the Upper Missouri River. Anthropocene 2, 51–64.
700
701 Stanley, E.H., Doyle, M.W., 2003. Trading off: the ecological effects of dam Removal.
702 Front. Ecol. Environ. 1(1), 15–22.
703
704 Suarez-Serrano, A., Alcaraz, C., Ibanez, C., Trobajo, R., Barata, C., 2010. Procambarus
705 clarkii as abioindicator of heavy metal pollution sources in the lower Ebro River and
706 Delta. Ecotoxicol. Environ. Saf. 73, 280–6.
707
708 Syvitski, J.P.M, Vörösmarty, C.J., Kettner, A.J., Green, P., 2005. Impact of humans on
709 the flux of terrestrial sediment to the global coastal ocean. Science 308, 376–80.
710
711 Tena, A., Ksiazek, L., Vericat, D., Batalla, R.J., 2012. Assessing the geomorphic effects
712 of a flushing flow in a large regulated river. River Res. Applic. 29, 876-890.
713
714 Turner, A., Millward, G.E., 2002. Suspended particles: their role in estuarine
715 biogeochemical cycles. Estuar. Coast. Shelf Sci. 55, 857– 883
716
31
717 Ullrich, S.M., Ilyushchenko, M.A., Kamberov, I.M., Tanton, T.W., 2007. Mercury
718 contamination in the vicinity of a derelict chlor-alkali plant. Part I: Sediment and water
719 contamination of Lake Balkyldak and the River Irtysh. Sci. Total Environ. 381, 1–16.
720
721 UNEP, United Nations Environment Programme Global Mercury Assessment Report.
722 2002. Geneva, Switzerland: UNEP Chemicals,
723 [http://www.unep.org/gc/gc22/Document/UNEP-GC22-INF3.pdf].
724
725 US EPA. Method 7473, mercury in solids and solutions by thermal decomposition,
726 amalgamation, and atomic absorption spectrophotometry; 2007.
727
728 Vericat, D., Batalla, R.J., 2006. Sediment transport in a large impounded river: the
729 lower Ebro, NE Iberian Peninsula. Geomorphology 79, 72–92.
730
731 Vörösmarty, C.J., Meybeck, M., Fekete, B., Sharma, K., Green, P., Syvitski, J.P.M.,
732 2003. Anthropogenic sediment retention: major global impact from registered river
733 impoundments. Glob. Planet Change 39,169–90.
734
735 Wildi, W., Dominik, J., Loizeau, J.L, Thomas, R.L., Favarger, P.Y., Haller, L., et al.,
736 2004. River, reservoir and lake sediment contamination by heavy metals downstream
737 from urban areas of Switzerland. Lakes Reserv. Res. Manage. 9, 75–87.
738
32
739 Wright, L.D., 1995. Morphodynamics of Inner Continental Shelves. CRC Press, Boca
740 Raton, 241pp.
741
742 Yeager, K.M., Schwehr, K.A., Louchouarn, P., Feagin, R.A., Schindler, K.J., Santschi,
743 P.H. 2018. Mercury inputs and redistribution in the Penobscot River and estuary,
744 Maine. Sci. Total Environ. 622–623, 172–183.
745
746 Yellen, B., Woodruff, J.D., Ralston, D.K., MacDonald, D.G., Jones, D.S., 2017. Salt
747 wedge dynamics lead to sediment trapping within side-embayments in high-energy
748 estuaries. J. Geophys. Res. Oceans http://dx.doi.org/10.1002/2016JC012595.
33
Reservoir Reservoir Mora Estuary Estuary Riba-roja Gracia River South North D’Ebre Island Mouth
Max 0.43 173.69 3.64 3.99 3.13 1.41
Min 0.07 7.24 0.75 0.83 0.62 0.66
Mean 0.19 41.62 1.61 1.93 1.01 0.85
Sd Dv 0.11 10.39 0,89 0.53 0.53 0.02
749
750 Table 1. Maximum, minimum, mean and standard deviation of the Hg concentrations
751 (in mg kg-1) in suspended particulate matter samples from Riba-Roja, from the Flix
752 Reservoir on the south bank (Reservoir South) where the industrial waste deposit was
753 accumulated (RS in Fig. 1C), from the Flix Reservoir on the north bank (Reservoir
754 North) opposite the industrial complex (RN in Fig. 1C), from Mora d’Ebre, from Gracia
755 Island and from the river mouth (Location in Fig. 1A).
756
34
Gracia Island River Mouth
May FF November FF May FF November FF
D. TMF 286.0 569.5 11.1 91.1 (gr m-2 d-1) Max. SPM conc. 298.0 557.5 6.7 184.7 (mg L-1) Cum. SPM Flux 3.74 8.8 0.05 0.08 (T m-2) D. Hg conc. 0.86 0.47 0.71 0.55 (mg kg-1) D. Hg Flux 246.8 268.9 7.9 50.6 (µg m-2 d-1) 757
758 Table 2: Downward total mass flux (D. TMF), maximum SPM concentration (Max.
759 SPM conc.), cumulative suspended particulate matter flux (Cum. SPM Flux), downward
760 Hg concentration (D. Hg conc.) and downward Hg flux (D. Hg Flux) during the
761 flushing flow sampling period.
35
Figure 1 Click here to download Figure: Fig 1 JEM last.pdf
A
Figure 1. A) Google Earth image of the lower Ebro River showing the locations where the benthic tripods were deployed (yellow stars), where the water samples were taken (orange triangles and yellow stars), and where the two gauging stations of the Ebro Water Authorities (Confederación Hidrográfica del Ebro) were installed (Ascó and Tortosa) (green circles). B) Google Earth image of the Flix Reservoir and the Flix meander before the extraction of the waste deposit. C) Bathymetry of the Flix Reservoir before the extraction of the waste deposit. RS: Reservoir South bank, where the industrial waste deposit (pale orange surface) has accumulated. RN: Reservoir North bank. HPP. Channels: underground hydraulic power plant channels. Orange triangles: water sampling stations. Yellow star: tripod position. Coordinates are in UTM, zone 31. Figure 2 Click here to download Figure: Fig 2 JEM last.pdf
D1 D2 2000 May FF Nov. FF 1600 )
-1 1200 Water Disch. (Ascó) s 3 Water Disch. (Tortosa)
(m 800
400
0
t r r y e y g p v c b l a c e p a n u e o e u O F M A u A S N D M J ------J 1 - 1 1 1 1 1 1 1 1 1 1 1600 1600 ) ) 1200 1200 -1 -1 s s 3 3 800 800 (m (m 400 400 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 : : : : : : : : : : : : : 6 2 8 0 6 2 8 6 2 8 0 6 2 0 1 1 0 0 1 1 0 1 1 0 0 1 Click heretodownloadFigure:Fig3JEMlast.pdf Figure 3 Reservoir (Flix R), Mora d’Ebre Figure 3: Hg concentrations in crosses). Standard deviation (green lines). the industrial waste deposit was complex (RN in Fig and 1C Reservoir North in Table 1), whereas rhombuses at the FlixReservoir Conc Hg (mg kg-1) 100 0.1 10 1 4 2 0 06 02 0 20 40 60 80 100 120 140
R-Roja
Flix R. Upstream DistancefromCoastline(km) the SPM sampled along the lower (M. Gracia d’Ebre), Island (Gra accumulated (RS in Fig 1C and correspond toHg values inthe
M. d'Ebre cia I.) and river mouth (R. Mouth) stations. White SPM fromthe north bank,opposite the industrial Reservoir South in Table 1). Mean values (red Ebro at River the Riba-roja the (R.roja), Flix black points are from the south bank, where
Gràcia I.
R. Mouth Figure 4 Click here to download Figure: Fig 4 JME last.pdf 1600
) 1200 Water Disch. (Ascó) -1 s
3 800
(m 400 0 SPM C (mg L 2.5 400 )
-1 Riba-roja 2 300 1.5 200 1 0.5 100 -1 Hg (mg kg 0 0 ) SPM C (mg L 2.5 400 )
-1 Flix Meander 2 300 1.5 200 1 0.5 100 -1 Hg (mg kg 0 0 ) SPM C (mg L 2.5 400 )
-1 Mora d'Ebre 2 300 1.5 200 1 0.5 100 -1 Hg (mg kg 0 0 )
1000 800 Water Disch. (Tortosa) ) -1 600 s 3 400 (m 200 0 SPM C (mg L 1 Gracia 600 ) -1 0.8 Island 0.6 400 0.4 200 0.2 -1 Hg (mg kg 0 0 )
0 0 0 0 0 0 0 0 0 0 0 0 : : : : : : 6 2 8 0 6 2 0 1 1 0 0 1 Hours, 21 November
Figure 4. Time series of water discharge at the Ascó and Tortosa gauging stations along with the Hg concentration and SPM concentration (SPM C) measured at Riba-roja, the Flix meander, Mora d’Ebre and Gracia Island during the November FF. Figure 5 Click here to download Figure: Fig 5 JME last.pdf FLIX RESERVOIR D1 FLIX RESERVOIR D2 1600 1200 Water discharge (m3 s-1) May FF November FF 800 400 0 18 16 Temperature ºC 14 12 10 360 Current direction (º) 270 180 90 0 40 30 Current speed (cm s-1) 20 10 0 1000 -1 750 SPM conc. 30-1000 (mg L ) 500 SPM conc. 0-30 (mg L-1) 250 30 0 400 300 Total Mass Flux (g m-2 d-1) 200 100 0 2-Dec 5-Dec 8-Dec 9-May 3-May 6-May 27-Apr 21-Apr 12-Apr 30-Apr 24-Apr 15-Apr 18-Apr 11-Dec 12-May 29-Nov 11-Nov 14-Nov 17-Nov 20-Nov 23-Nov 26-Nov
Figure 5: Time series of water discharge at the Ascó gauging station along with temperature, current direction, current speed, SPM concentration (SPM conc.) and downward total mass flux recorded in the Flix Reservoir during deployments 1 and 2. Note the two different scales of SPM concentration. D1: deployment 1. D2: deployment 2. FF: flushing flow Figure 6 Click here to download Figure: Fig 6 JME last.pdf GRACIA ISLAND D1 GRACIA ISLAND D2 1000 750 Water Discharge (m3 s-1) May FF November FF 500 250 0 40 30 Salinity 20 10 0 360 270 180 Current Direction 90 (º) 0 75 Current Speed (cm s-1) 50
25
0 600 SPM conc. 10-1000 (mg L-1) 400 SPM conc. 0-10 (mg L-1) 200 10 0 600
400 Total Mass Flux (g m-2 d-1)
200
0 2-Dec 5-Dec 8-Dec 6-May 9-May 3-May 27-Apr 30-Apr 21-Apr 24-Apr 12-Apr 15-Apr 18-Apr 14-Nov 17-Nov 20-Nov 23-Nov 11-Nov 12-May 11-Dec 26-Nov 29-Nov Figure 6: Time series of water discharge at the Tortosa gauging station along with salinity, current direction, current speed, SPM concentration (SPM conc.) and downward total mass flux recorded at the Gracia Island station during deployments 1 and 2. Note the two different scales of SPM concentration. D1: deployment 1. D2: deployment 2. FF: flushing flow. Figure 7 Click here to download Figure: Fig 7 JME last.pdf RIVER MOUTH D1 RIVER MOUTH D2 1000 750 Water Discharge (m3 s-1) May FF November FF 500 250 0 40 30 Salinity 20 10 0 360 Current Direction (º) 270 180 90 0 40 Current Speed (cm s-1) 30 20 10 0 200 -1 150 SPM conc. 10-1000 (mg L ) -1 100 SPM conc. 0-10 (mg L ) 50 10 0 400 300 Total Mass Flux (g m-2 d-1) 200 100 0 2-Dec 5-Dec 8-Dec 3-May 6-May 9-May 24-Apr 27-Apr 30-Apr 15-Apr 18-Apr 21-Apr 12-Apr 11-Dec 29-Nov 11-Nov 14-Nov 17-Nov 20-Nov 23-Nov 26-Nov 12-May Figure 7: Time series of water discharge at the Tortosa gauging station, salinity, current direction, current speed, SPM concentration (SPM conc.) and downward total mass flux recorded at the river mouth station during deployments 1 and 2. Note the two different scales of SPM concentration. D1: deployment 1. D2: deployment 2. FF: flushing flow Figure 8 Click here to download Figure: Fig 8 JME last.pdf
Deployment 1 Deployment 2 May FF Nov FF
5000 24 5000 24 Hg conc.(mgkg ) Flix Reservoir -1 4000 20 4000 20 d -2 16 16 3000 3000 Flix Reservoir 12 12 2000 2000 8 8
1000 1000 -1
4 4 ) Hg flux (µg m 0 0 0 0
300 1 300 1 Hg conc.(mg kg ) -1 0.8 Gracia Island 0.8 d -2 200 200 Gracia Island 0.6 0.6 0.4 0.4 100 100
0.2 0.2 -1 ) Hg flux (µg m 0 0 0 0
300 1 300 1 Hg conc.(mg kg ) -1 0.8 River Mouth 0.8 d -2 200 200 0.6 0.6 River Mouth 0.4 0.4 100 100
0.2 0.2 -1 ) Hg flux (µg m 0 0 0 0 2-Dec 5-Dec 8-Dec 6-May 9-May 3-May 17-Nov 20-Nov 11-Dec 23-Nov 11-Nov 26-Nov 29-Nov 14-Nov 21-Apr 24-Apr 12-Apr 27-Apr 15-Apr 30-Apr 18-Apr 12-May Date Date
Figure 8: Time series of Hg concentration and Hg fluxes in the trap-collected particles of the Flix Reservoir, Gracia Island and the river mouth during the two deployment periods. FF: flushing flow. Supplementary Material Figure. 1 Click here to download Supplementary Interactive Plot Data (CSV): Supplementary material Fig. 1 JEM last.pdf
Supplementary Material Figure 2 Click here to download Supplementary Interactive Plot Data (CSV): Supplementary material Fig. 2 JEM last.pdf
Supplementary Material Figure 3 Click here to download Supplementary Interactive Plot Data (CSV): Supplementary material Fig. 3 JEM last.pdf